How to Deploy ComfyUI on Vultr Cloud GPU

Introduction

ComfyUI is an open-source, node-based application that lets you design, test, and execute diffusion model pipelines using a graphical web interface. ComfyUI runs advanced stable diffusion pipelines using a flowchart interface of interconnected nodes that support checkpoint loading, workflow re-executions, and generation from PNG files.

This article explains how to deploy ComfyUI on a Vultr Cloud GPU server. You will set up all project dependencies, install the ComfyUI Manager to manage models and run ComfyUI as a system service to run diffusion model pipelines on the server.

Prerequisites

Before you begin:

• Deploy an NVIDIA A100 Vultr Cloud GPU instance running Ubuntu.
• Set up a new domain A record pointing to the server's IP address. For example, comfyui.example.com.
• Access the server using SSH as a non-root sudo user. For example, linuxuser.
• Update the server.

Install ComfyUI

Follow the steps below to install ComfyUI in your working directory on the server.

1. Switch to your user home directory.

   CONSOLE

   Copy

   $ cd

   Explain Code

   Powered by Vultr AgentBeta

2. Clone the ComfyUI project repository using Git.

   CONSOLE

   Copy

   $ git clone https://github.com/comfyanonymous/ComfyUI.git

   Explain Code

   Powered by Vultr AgentBeta

3. Switch to the ComfyUI directory.

   CONSOLE

   Copy

   $ cd ComfyUI

   Explain Code

   Powered by Vultr AgentBeta

4. Install the required PyTorch, numpy <2, and xformers dependency packages.

   CONSOLE

   Copy

   $ sudo pip install "numpy<2" torch==2.6.0 torchvision==0.21.0 torchaudio==2.6.0 xformers==0.0.29.post3 --extra-index-url https://download.pytorch.org/whl/cu124

   Explain Code

   Powered by Vultr AgentBeta

   The above command installs the following packages:

   • torch: A machine learning library that provides a flexible and dynamic computational graph for deep learning tasks.
   • torchvision: A PyTorch library used for computer vision tasks. It includes utilities for image and video datasets that handle image classification and object detection tasks.
   • torchaudio: A PyTorch library used for audio processing tasks. It provides audio loading, transformations, and common audio datasets for deep learning applications.
   • xformers: A Python library for implementing transformer-based models and attention mechanisms.

5. Install additional ComfyUI dependencies using the requirements.txt file.

   CONSOLE

   Copy

   $ sudo pip install -r requirements.txt

   Explain Code

   Powered by Vultr AgentBeta

6. Switch to the models/checkpoints directory.

   CONSOLE

   Copy

   $ cd models/checkpoints

   Explain Code

   Powered by Vultr AgentBeta

7. Open the put_checkpoints_here file using a text editor such as nano.

   CONSOLE

   Copy

   $ nano put_checkpoints_here

   Explain Code

   Powered by Vultr AgentBeta

8. Add the following contents to the file.

   SH

   Copy

   # Checkpoints
   
   ### SDXL
   
   wget -c https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0/resolve/main/sd_xl_base_1.0.safetensors -P ./
   #wget -c https://huggingface.co/stabilityai/stable-diffusion-xl-refiner-1.0/resolve/main/sd_xl_refiner_1.0.safetensors -P ./
   
   # SD1.5
   
   wget -c https://huggingface.co/runwayml/stable-diffusion-v1-5/resolve/main/v1-5-pruned-emaonly.ckpt -P ./
   
   # SD2
   
   #wget -c https://huggingface.co/stabilityai/stable-diffusion-2-1-base/resolve/main/v2-1_512-ema-pruned.safetensors -P ./
   #wget -c https://huggingface.co/stabilityai/stable-diffusion-2-1/resolve/main/v2-1_768-ema-pruned.safetensors -P ./
   

   Explain Code

   Powered by Vultr AgentBeta

   Save and close the file.

   The above configuration downloads the Stable Diffusion XL and Stable Diffusion 1.5 models to the ComfyUI models directory using the wget utility. Modify the file to include new model checkpoint URLs from sources such as Hugging Face to enable in ComfyUI.

9. Run the file using bash to download the specified models.

   CONSOLE

   Copy

   $ bash put_checkpoints_here

   Explain Code

   Powered by Vultr AgentBeta

   Output:

   --2026-08-04 12:00:01--  https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0/resolve/main/sd_xl_base_1.0.safetensors
   Resolving huggingface.co (huggingface.co)... 3.163.189.90
   Connecting to huggingface.co (huggingface.co)|3.163.189.90|:443... connected.
   HTTP request sent, awaiting response... 200 OK
   Length: 6938040706 (6.5G) [binary/octet-stream]
   Saving to: './sd_xl_base_1.0.safetensors'
   
   sd_xl_base_1.0.safetensors   100%[==============================>]   6.46G   255MB/s    in 26s
   
   2026-08-04 12:00:27 (255 MB/s) - './sd_xl_base_1.0.safetensors' saved [6938040706/6938040706]
   
   --2026-08-04 12:00:27--  https://huggingface.co/runwayml/stable-diffusion-v1-5/resolve/main/v1-5-pruned-emaonly.ckpt
   Resolving huggingface.co (huggingface.co)... 3.163.189.90
   Connecting to huggingface.co (huggingface.co)|3.163.189.90|:443... connected.
   HTTP request sent, awaiting response... 200 OK
   Length: 4265380512 (4.0G) [binary/octet-stream]
   Saving to: './v1-5-pruned-emaonly.ckpt'
   
   v1-5-pruned-emaonly.ckpt     100%[==============================>]   3.97G   248MB/s    in 17s
   
   2026-08-04 12:00:44 (245 MB/s) - './v1-5-pruned-emaonly.ckpt' saved [4265380512/4265380512]

10. Switch to your user home directory.

    CONSOLE

    Copy

    $ cd

    Explain Code

    Powered by Vultr AgentBeta

11. Move the ComfyUI project directory to a system-wide location such as /opt.

    CONSOLE

    Copy

    $ sudo mv ComfyUI /opt

    Explain Code

    Powered by Vultr AgentBeta

Set Up ComfyUI as a System Service

Follow the steps below to set up ComfyUI as a system service to automatically run and manage the application's processes on your server.

1. Create a new comfyui.service system service file.

   CONSOLE

   Copy

   $ sudo nano /etc/systemd/system/comfyui.service

   Explain Code

   Powered by Vultr AgentBeta

2. Add the following configurations to the file. Replace /opt/ComfyUI with your actual ComfyUI project directory path.

   INI

   Copy

   [Unit]
   Description=ComfyUI System Service
   After=network.target
   
   [Service]
   WorkingDirectory=/opt/ComfyUI
   ExecStart=python3 main.py
   
   [Install]
   WantedBy=multi-user.target

   Explain Code

   Powered by Vultr AgentBeta

   Save and close the file.

   The above configuration creates a new comfyui system service to manage the ComfyUI application processes on the server.

3. Restart the systemd daemon to apply the new service changes.

   CONSOLE

   Copy

   $ sudo systemctl daemon-reload

   Explain Code

   Powered by Vultr AgentBeta

4. Enable the ComfyUI system service.

   CONSOLE

   Copy

   $ sudo systemctl enable comfyui

   Explain Code

   Powered by Vultr AgentBeta

5. Start the ComfyUI service.

   CONSOLE

   Copy

   $ sudo systemctl start comfyui

   Explain Code

   Powered by Vultr AgentBeta

6. View the ComfyUI service status and verify that it's active.

   CONSOLE

   Copy

   $ sudo systemctl status comfyui

   Explain Code

   Powered by Vultr AgentBeta

Configure Nginx as a Reverse Proxy to Access ComfyUI

ComfyUI listens for connection requests on the localhost port 8188 by default unless you set a specific IP address and port using the --listen option. Set up a reverse proxy such as Nginx to forward all incoming connection requests to the backend port 8188 to enable access to the ComfyUI interface. Follow the steps below to install and configure Nginx to forward HTTP requests to the ComfyUI interface using your comfyui.example.com domain.

1. Install Nginx.

   CONSOLE

   Copy

   $ sudo apt install nginx -y

   Explain Code

   Powered by Vultr AgentBeta

2. Create a new comfyui.conf Nginx virtual host configuration in the sites-available directory.

   CONSOLE

   Copy

   $ sudo nano /etc/nginx/sites-available/comfyui.conf

   Explain Code

   Powered by Vultr AgentBeta

3. Add the following configurations to the file. Replace comfyui.example.com with your actual domain.

   NGINX

   Copy

   upstream backend {
       server 127.0.0.1:8188;
   }
   
   server {
   
       listen 80;
       listen [::]:80;
   
       server_name comfyui.example.com;
   
       proxy_set_header Host $host;
       proxy_http_version 1.1;
       proxy_set_header X-Real-IP $remote_addr;
       proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
       proxy_set_header X-Forwarded-Proto $scheme;
       proxy_set_header Sec-WebSocket-Extensions $http_sec_websocket_extensions;
       proxy_set_header Sec-WebSocket-Key $http_sec_websocket_key;
       proxy_set_header Sec-WebSocket-Version $http_sec_websocket_version;
       proxy_set_header Upgrade $http_upgrade;
       proxy_set_header Connection "Upgrade";
   
       location / {
           proxy_pass http://backend$request_uri;
       }
   }

   Explain Code

   Powered by Vultr AgentBeta

   Save and close the file.

4. Link the configuration file to the sites-enabled directory to activate it.

   CONSOLE

   Copy

   $ sudo ln -s /etc/nginx/sites-available/comfyui.conf /etc/nginx/sites-enabled/

   Explain Code

   Powered by Vultr AgentBeta

5. Test the Nginx configuration for errors.

   CONSOLE

   Copy

   $ sudo nginx -t

   Explain Code

   Powered by Vultr AgentBeta

   Output:

   nginx: the configuration file /etc/nginx/nginx.conf syntax is ok
   nginx: configuration file /etc/nginx/nginx.conf test is successful

6. Restart Nginx to apply the configuration changes.

   CONSOLE

   Copy

   $ sudo systemctl restart nginx

   Explain Code

   Powered by Vultr AgentBeta

Security

Uncomplicated Firewall (UFW) is active on Vultr Ubuntu servers by default. Follow the steps below to enable connections to the HTTP port 80, and the HTTPS port 443 through the firewall configuration to enable access to the ComfyUI interface.

1. View the UFW status to verify that the firewall is active on your server.

   CONSOLE

   Copy

   $ sudo ufw status

   Explain Code

   Powered by Vultr AgentBeta

   Allow SSH connections on port 22 and enable UFW if the firewall status is inactive.

   CONSOLE

   Copy

   $ sudo ufw allow 22/tcp && sudo ufw enable 

   Explain Code

   Powered by Vultr AgentBeta

2. Allow network connections to the HTTP port 80.

   CONSOLE

   Copy

   $ sudo ufw allow 80/tcp

   Explain Code

   Powered by Vultr AgentBeta

3. Allow connections to the HTTPS port 443.

   CONSOLE

   Copy

   $ sudo ufw allow 443/tcp

   Explain Code

   Powered by Vultr AgentBeta

4. Reload the UFW configuration to apply the firewall changes.

   CONSOLE

   Copy

   $ sudo ufw reload

   Explain Code

   Powered by Vultr AgentBeta

Secure ComfyUI with Let's Encrypt SSL Certificates

SSL certificates encrypt network connections between a user's web browser and the server. Generate SSL certificates using a trusted certificate authority such as Let's Encrypt to secure network connections to the server using HTTPS. Follow the steps to install the Certbot Let's Encrypt client and generate trusted SSL certificates to secure the ComfyUI interface.

1. Install the Certbot Let's Encrypt client tool using Snap.

   CONSOLE

   Copy

   $ sudo snap install --classic certbot

   Explain Code

   Powered by Vultr AgentBeta

2. Generate a new SSL certificate using your domain. Replace comfyui.example.com and user@example.com with your actual domain and active email address respectively.

   CONSOLE

   Copy

   $ sudo certbot --nginx -d comfyui.example.com -m user@example.com --agree-tos

   Explain Code

   Powered by Vultr AgentBeta

3. Test the Certbot automatic SSL certificate renewal process.

   CONSOLE

   Copy

   $ sudo certbot renew --dry-run

   Explain Code

   Powered by Vultr AgentBeta

   Certbot auto-renews your SSL certificate every 90 days before expiry if the above command is successful.

Generate Images using ComfyUI

ComfyUI is actively running as a system service on your server and accessible using the comfyui.example.com domain defined in your Nginx configuration. Follow the steps below to set up the default nodes and generate images using ComfyUI.

1. Visit your domain using a web browser such as Chrome to access the ComfyUI interface.

   https://comfyui.example.com

   

   ComfyUI uses the text-to-image workflow by default. Click Load Default in the floating menu to switch to the default workflow in case of unintended changes.

2. Click the Load Checkpoint node dropdown to select a diffusion model. For example, select the Stable Diffusion XL model.

   

3. Navigate to the Prompt node, and enter a positive prompt such as beautiful scenery nature glass bottle landscape, purple galaxy bottle,. Enter a negative prompt in the respective node such as text,watermark.

   

4. Click Queue Prompt on the main floating menu to generate a new image.

5. Wait for the image generation process to complete and view your generated image in the Save Image node.

   

Enable ComfyUI Manager

ComfyUI manager is an add-on tool that provides a user-friendly interface to install and manage custom nodes or models. Follow the steps below to install and enable the ComfyUI manager tool.

1. Switch to the ComfyUI project directory.

   CONSOLE

   Copy

   $ cd /opt/ComfyUI

   Explain Code

   Powered by Vultr AgentBeta

2. Stop the ComfyUI system service.

   CONSOLE

   Copy

   $ sudo systemctl stop comfyui

   Explain Code

   Powered by Vultr AgentBeta

3. Switch to the custom_nodes directory.

   CONSOLE

   Copy

   $ cd custom_nodes

   Explain Code

   Powered by Vultr AgentBeta

4. Clone the ComfyUI manager repository using Git.

   CONSOLE

   Copy

   $ git clone https://github.com/ltdrdata/ComfyUI-Manager.git

   Explain Code

   Powered by Vultr AgentBeta

5. Start the ComfyUI system service.

   CONSOLE

   Copy

   $ sudo systemctl start comfyui

   Explain Code

   Powered by Vultr AgentBeta

6. Access the ComfyUI interface in your web browser.

   https://comfyui.example.com

7. Verify and click the new Manager option on the floating menu.

   

8. Verify that the ComfyUI manager interface loads correctly.

   

9. Click Install Custom Nodes or Install Models to install additional nodes and models to enable in ComfyUI.

   

Common ComfyUI Nodes

ComfyUI nodes are modular units that form a workflow and perform specific image generation functionalities. Each ComfyUI node processes input data and passes the output to the next node for further processing in a workflow. This modular approach enables you to create diverse image generation workflows and test multiple configurations to generate high-quality results. The following are the most common node types you can use in your ComfyUI workflows.

• Input nodes: Export data to other nodes, such as images, text prompts, and random numbers.
• Processing nodes: Manipulate the data provided by input nodes, such as resizing images, applying filters, and generating prompts.
• Output nodes: Save results from the image generation process, such as saving images to disk or displaying them in a preview window.
• Load Checkpoint: Loads a pre-trained diffusion model in ComfyUI.
• CLIP Encode: Encodes text prompts into CLIP embeddings.
• Prompt Node: Provides text prompts to the model.
• Negative Prompt Node: Provides negative prompts to the model to guide the image generation process.
• VAE Encode: Converts images to latent representations.
• VAE Decode: Converts latent representations to images.
• Random Seed Node: Generates random seeds for Image generation samplers.
• Empty Latent Image Node: Creates a blank latent image of a specified size for use in other image generation nodes.
• Scale Node: Scales images to a desired size.
• Save image Node: Saves generated images to disk.
• Preview Node: Displays generated images in a preview window.

Conclusion

You have deployed ComfyUI on a Vultr Cloud GPU server and generated images using the ComfyUI web interface. In addition, you installed the ComfyUI Manager to enable new nodes and models. For more information and configuration options, please visit the ComfyUI community documentation.